Les failles actives

Identification - images satellitaires - terrain

Fonctionnement - vitesse moyenne - cycle sismique

124!W

124!W

122!W

122!W

120!W

120!W

118!W

118!W

116!W

116!W

32!N 32!N

34!N 34!N

36!N 36!N

38!N 38!N

40!N 40!N

Sismicité de la Californie (1973-2004)

La faille de San Andreas n'est pas active ???

Ruptures de surface du séisme de Kokoxili (2001, Mw = 7.9)

Une faille active modifie le paysage

Combinaison d’images SPOT (10 m) et IKONOS (! 1 m)

(Documents Y. Klinger, Tectonique IPGP)

Faille du Kunlun

Formation des facettes triangulaires

Versant ouest contrôlé par une faille Versant est contrôlé par l'érosion

Tanghenanshan

Topographienumérique

Image Landsat

Tanghenanshan

Topographie + Image satellitaire

SW NE

© Yves Gaudemer, 2005

SW NE

Van der Woerd et al. 2003

Plis actifs

Meyer 1991

Shibaocheng

Photo P. Tapponnier

120 km

Décalage du Fleuve Jaune (Huanhe) par la

faille de Haiyuan

Comment fonctionne une faille ?

Si la faille fonctionne régulièrement, avec un séisme caractéristique, le temps de

récurrence T de celui-ci est simplement :T = U/V

où U est le déplacement cosismiqueet V est la vitesse de glissement sur la faille.

Exemple :M = 8, U = 10 m, V = 1 cm/an —> T = 1000 ans

Déterminer U et V

At six different sites we dated as many as 93 quartz pebblesand 22 organic samples to constrain the ages of the cumulativeoffsets of 11 distinct morphological markers. The large numberof samples is paramount in obtaining statistically meaningfulcosmogenic ages, and in identifying clearly and discardingoutliers on any given terrace. Several independent values of

both the marker ages and slip-rates, based either on cosmo-genic or 14C dating, are consistent with one another at each siteand from site-to-site (Fig. 31), giving us confidence in individualrate determinations. Overall, our estimate of the slip-rate isdetermined within an uncertainty of t2.0 mm yrx1, and theaverage value of 11.5 mm yrx1 is robust (Fig. 31).For offset risers, the ages we take, and hence the slip-rates,

depend on field interpretation regarding the terraces (strathin general, fill in one case), the nature of which may vary fromsite-to-site or at a given site. An end-member scenario, con-sidered highly unlikely, is that the ages of all riser offsets arethose of upper terraces. Slip-rates derived in this scenarioare y7 mm yrx1. In view of geological evidence, we believe7 mm yrx1 to be implausibly small. Nonetheless, this value isgreater than the rate inferred from two epoch remeasurementsof the few extant GPS stations north and south of the KunlunFault (Chen et al. 2000). Until a denser and longer GPS surveyis undertaken, slip-rates based on this technique should beregarded as preliminary.Our average rate value of 11.6t0.8 mm yrx1 in Xidatan–

Dongdatan (Fig. 15), which is based only on cosmogenicdating of offset terrace risers, is consistent with those derivedfrom other geological studies with different dating tech-niques. Zhao et al. (1994) and Zhao (1996), in particular,documented 10 offsets of gullies and terrace risers in Xidatan,ranging from 10 to 152 m, and retrieved seven TL or 14Cages, ranging from 2.14 to 12.0 kyr BP, with uncertainties of6–9 per cent. From this, they derived a Holocene slip-rateof about 11.5 mm yrx1. Unfortunately, without error estimateson the offsets, the uncertainty on that rate cannot be assessed.The average Quaternary rate estimated by Kidd & Molnar(1988) (10–20 mm yrx1), loosely brackets both our values andthose of Zhao et al. In a strict sense, however, it cannot besimply compared with either, because it corresponds to a muchlonger time-span. Recall that the long-term rate of Kidd &Molnar (1988) is deduced from the separation between LowerPleistocene moraines containing boulders of pyroxenite, southand east of the Kunlun Pass, and their presumed source area,30 km to the west, in the mountains north of the fault. Thefactor of two in the estimate comes chiefly from the largeuncertainty in the age of the lake-beds that overlie the moraine(2.8–1.5 Ma) (Kidd & Molnar 1988; Qian et al. 1982; Wu et al.1982). To this uncertainty on the age, one should also probablyadd a y30 per cent error on the offset, due to poor definitionof the eastern piercing point (Kidd & Molnar 1988: Fig. 6).Hence, while it is interesting to note that there is no first-order discrepancy between the loosely constrained, averageQuaternary rate (Kidd & Molnar 1988) and the Holocene slip-rates derived from 14C, TL or cosmogenic dating, kinematicmodels of the current tectonics of northeast Tibet making useof the latter rates will be more precise.

5.2 Climatic origin of the morphology

Along the Xidatan–Dongdatan valleys, the relatively smallnumber of terrace levels (about seven, Fig. 5) found along mostof the streams implies a relatively uniform morphologicalresponse on a regional scale. It seems that most of the streamsemplaced terraces at discrete, comparable times. Not all thestreams deposited all the terrace levels identified, however. Thevariable distribution of terraces along the different streamsis thus probably due to local conditions in each watershed.

Figure 28. View, looking west, along fault strike of the shutter ridgewest of Maqen (see Fig. 27b for location). The stream incision acrossthe ridge reveals several tens of metres of thick, vertical gouge.

Figure 29. View of the offset LGMmoraine ridge and T3 terrace level,abandoned after 11 156t157 yr BP.

Figure 30. View, looking south, of a small gully offset by the KunlunFault. The offset, probably resulting from only one large earthquake, isabout 12 m.

382 J. Van Der Woerd et al.

# 2002 RAS, GJI 148, 356–388

Mesure du déplacement cosismique U

<— Faille du Kunlun (Tibet)

12 mSéisme de Manyi (M = 7.6, Tibet)

—>

(Van der Woerd et al., 2002)

(Peltzer et al., 1999)

(Peltzer et al., 1999)

<—

Pour calculer V, on mesure le décalage d'objets (moraines, cônes alluviaux, etc) que l'on peut dater (14C, 10Be, 26Al, 37Cl)

|||

|

|

||

|

||||

|||||

|||

|||

||

||

||

||||

|||

|||||||||

||||

|

||

|

|||

||||||||

|

||||

||||||||||||||

|||||||||||

||||||||||||

||||||||||

|

|

|||||||||||

|||||||||||||

||||||||

|||||||||||||

||||

|||||

|||||

|

||||||||||||||||||

|||||||||||||||||||||

|||||||||||||

|

||||||||||||

WG EG

WV

CO

EOWO

M2

2 km

M1 EC

63±11ka

63±9 ka

61±9 ka

63±7 ka 61±12 ka

77±15 ka

72±13 ka

51±9 ka

53±5 ka

38±4 ka

40±6 ka

46±8 ka

35±5 ka

32±6 ka

84±16 ka

22±5 ka

41±9 ka

32±5 ka

52±7 ka

6±2 ka

43±4 ka

47±4 ka

40±5 ka

49±7 ka

44±5ka

105±13 ka

107±11 ka

53±6 ka

53±6 ka

85±11 ka 12±2 ka

30±3 ka

22±3 ka

35±5 ka

14±4 ka

17±3ka

4±1 ka

70±10 ka

41±6 ka

73±8 ka

42±8 ka

(Mériaux et al., 2000)

26.3 ± 1.4 mm/an

(Mériaux et al., 2000)

At six different sites we dated as many as 93 quartz pebblesand 22 organic samples to constrain the ages of the cumulativeoffsets of 11 distinct morphological markers. The large numberof samples is paramount in obtaining statistically meaningfulcosmogenic ages, and in identifying clearly and discardingoutliers on any given terrace. Several independent values of

both the marker ages and slip-rates, based either on cosmo-genic or 14C dating, are consistent with one another at each siteand from site-to-site (Fig. 31), giving us confidence in individualrate determinations. Overall, our estimate of the slip-rate isdetermined within an uncertainty of t2.0 mm yrx1, and theaverage value of 11.5 mm yrx1 is robust (Fig. 31).For offset risers, the ages we take, and hence the slip-rates,

depend on field interpretation regarding the terraces (strathin general, fill in one case), the nature of which may vary fromsite-to-site or at a given site. An end-member scenario, con-sidered highly unlikely, is that the ages of all riser offsets arethose of upper terraces. Slip-rates derived in this scenarioare y7 mm yrx1. In view of geological evidence, we believe7 mm yrx1 to be implausibly small. Nonetheless, this value isgreater than the rate inferred from two epoch remeasurementsof the few extant GPS stations north and south of the KunlunFault (Chen et al. 2000). Until a denser and longer GPS surveyis undertaken, slip-rates based on this technique should beregarded as preliminary.Our average rate value of 11.6t0.8 mm yrx1 in Xidatan–

Dongdatan (Fig. 15), which is based only on cosmogenicdating of offset terrace risers, is consistent with those derivedfrom other geological studies with different dating tech-niques. Zhao et al. (1994) and Zhao (1996), in particular,documented 10 offsets of gullies and terrace risers in Xidatan,ranging from 10 to 152 m, and retrieved seven TL or 14Cages, ranging from 2.14 to 12.0 kyr BP, with uncertainties of6–9 per cent. From this, they derived a Holocene slip-rateof about 11.5 mm yrx1. Unfortunately, without error estimateson the offsets, the uncertainty on that rate cannot be assessed.The average Quaternary rate estimated by Kidd & Molnar(1988) (10–20 mm yrx1), loosely brackets both our values andthose of Zhao et al. In a strict sense, however, it cannot besimply compared with either, because it corresponds to a muchlonger time-span. Recall that the long-term rate of Kidd &Molnar (1988) is deduced from the separation between LowerPleistocene moraines containing boulders of pyroxenite, southand east of the Kunlun Pass, and their presumed source area,30 km to the west, in the mountains north of the fault. Thefactor of two in the estimate comes chiefly from the largeuncertainty in the age of the lake-beds that overlie the moraine(2.8–1.5 Ma) (Kidd & Molnar 1988; Qian et al. 1982; Wu et al.1982). To this uncertainty on the age, one should also probablyadd a y30 per cent error on the offset, due to poor definitionof the eastern piercing point (Kidd & Molnar 1988: Fig. 6).Hence, while it is interesting to note that there is no first-order discrepancy between the loosely constrained, averageQuaternary rate (Kidd & Molnar 1988) and the Holocene slip-rates derived from 14C, TL or cosmogenic dating, kinematicmodels of the current tectonics of northeast Tibet making useof the latter rates will be more precise.

5.2 Climatic origin of the morphology

Along the Xidatan–Dongdatan valleys, the relatively smallnumber of terrace levels (about seven, Fig. 5) found along mostof the streams implies a relatively uniform morphologicalresponse on a regional scale. It seems that most of the streamsemplaced terraces at discrete, comparable times. Not all thestreams deposited all the terrace levels identified, however. Thevariable distribution of terraces along the different streamsis thus probably due to local conditions in each watershed.

Figure 28. View, looking west, along fault strike of the shutter ridgewest of Maqen (see Fig. 27b for location). The stream incision acrossthe ridge reveals several tens of metres of thick, vertical gouge.

Figure 29. View of the offset LGMmoraine ridge and T3 terrace level,abandoned after 11 156t157 yr BP.

Figure 30. View, looking south, of a small gully offset by the KunlunFault. The offset, probably resulting from only one large earthquake, isabout 12 m.

382 J. Van Der Woerd et al.

# 2002 RAS, GJI 148, 356–388

Paléo-sismologie

LichenométrieDendrochronologie

GéomorphologieTranchée de l'Evêque (Yammoûneh, 2002 – CNRSL & IPGP)

Photo J. Vanderwoerd

Quand a eu lieu le dernier séisme ?

Remblai

02

03a

03c

03d

03f03e

03c

05

04d

04a04b 04b05

03b

01

06a 07a 07a

07a

07a

07b

04a

04c

03d/f

06b08

08

08

05

04+05

06c

10

10

0909

1011

11

1113

15

15

1412

14

14

12

15

12

13

14

15

16

1617a

17a

(17/18)b 19b

20

23

23

24-25

21

2219/21?18a

27 28

27

29

26

27

28

29

30

3132

34

35

37

39a

3840a4142

4546

4849

50

4344

36

33

30

28

29

30

23

22

23

27

23

28

29

30

27

28

29

30 31

32

33

34

35

3637

38 39b40b

41

39b-40b

42+43

44

4546

47

47

{

54

55

(56+57)?

{

50{

54

55

56

57

58

59

60

61

{

48

49

5152

53

53

51

52

F1

F2

F4

F5

F6

F7

F3

F10F9

F8

C3C2

C1

C4

C5

Kazzâb Trench

(Yammoûneh basinIPGP / CNRSL, 2001)

1 m

E W

Tranchée de Kazzab (IPGP/CNRSL, 2001)

Daëron 2004

Temps Temps Temps

Glis

sem

ent

!max

!min

Parfaitement périodique T prévisibleTn = f(Un-1)

U prévisibleUn = f(Tn-1)

U = VT

U = VT U =

VT

Trois modèles de séismes pour une failleà vitesse moyenne constante

4000

10

20

30

40

50

800 1200 1600 2000

Année

Dép

lace

men

t cu

mul

é (m

)

8.9 cm/an

2.4 cm/an3.1 cm/an

Les failles fonctionnent-elles vraiment à vitesse constante ?Sur quelle échelle de temps ?

(d'après Weldon et al. 2004)

Faille de San Andreas (Wrightwood, Californie)

Question :les failles fonctionnent-elles régulièrement ?

Certains physiciens et sismologues pensent que le fonctionnement des failles est complètement aléatoire

1

0 2 4 6 8

10

100

1000

10000

100000

b ! 1

Magnitude

Nom

bre

de s

éis

mes

Loi de Gutenberg-Richter

Log N(M) = a - bM

Modèle de patins et ressorts

Dessin : Y. Gaudemer

Toute tentative de prédiction/prévision serait vaine...

Critère de Mohr-CoulombÀ la rupture, la contrainte cisaillante " et la contrainte

normale !n sur les surfaces potentielles de rupture sont liées par une relation linéaire :

" = co + #!n

l'enveloppe de Mohr est une droite

"

!n

co

$ # = tan($)

" = c o + #

! n

co est la cohésion# est le coefficient de friction interne

$ est l'angle de friction interne(analogie avec le frottement)

Contrainte de Coulomb

!n

"

" = c o

+ #! n

#!n

"max

" = c o

+ #! n

#!n

!C = " - #!n

Il est facile de voir que la quantité !C = " - #!n

(ou !C = " - #(!n - pf) s'il y a des fluides)ne peut jamais excéder une valeur critique. Cette quantité !C est appelée contrainte de Coulomb.

Quelle est la relation entre co et la valeur critique ?

!C = "max - #!n

contrainte normale contrainte normale

cont

rain

te c

isai

llant

e

cont

rain

te c

isai

llant

e

La notion de cycle sismique

! Juste après le séisme précédent

" !C augmente mais la faille reste

bloquée

# !C atteint la valeur critique :

c'est le séisme

$ Les contraintes sont relâchées, un nouveau cycle commence...

Il est très difficile de mesurer la contrainte de Coulomb et, donc, de savoir si une faille est proche de la rupture ou non.

Mais on peut calculer facilement les variations de !C engendrées par des séismes proches.

+ =

Rise DropA. Coulomb stress change for right-lateral faults parallel to master fault Stress

B. Coulomb stress change for faults optimally oriented for failure

N27°E regional compression (!r) of 100 bars; µ’ = 0.75

!r

right-lateral shearstress change

effective friction x

normal stress change

right-lateral Coulombstress change

+ =

µ' (-!n)"s !f+ =

+

shear stresschange

effective friction x

normal stress change

Coulomb stresschange

+ =

µ' (-!n)"s !f+ =

+

left-lateral

right-lateralOptimum

Slip Planes

R R

opt

+ =

%!C = %" + #%!n

%" #%!n %!C

Exemple d'un séisme sur une faille dextre :

augmentation diminutionsans changement

0 20 40 600

20

40

60

80

100

Optimum right-lateral planes

JoshuaTree

Homestead

Galway Lake

Landers

Valley

Coulomb Stress Change (bars)

-.50-.250

.50

.75

-.75

.25

SpringsNorth Palm

km

1992 Landersrupture trace

Exemple : le séisme de Landers (Californie, 1992, M = 7.3)

Les séismes de Galway Lake (1975), Homestead Valley (1979), North Palm Spring

(1986) et Joshua Tree (1992) ont modifié !C dans la

région du futur séisme de Landers.

Conséquences sur la faille de San Andreas

33°30'

116°

117°3

0'

35°30'

Coulomb Stress

Change (bars)

25 km

ML!1 during 25 days

after Landers main shock

-0.2-0.10.0

0.20.3

-0.3

0.1

S A

N

J A C

I N T O

S A N

A N

D R

E A

S

27 Nov-4 Dec

Mojave SegmentCoachella Valley

SegmentSan Bernardino Mtn. Segment

35 mm/yr, last event in 1857 24±3 mm/yr, 1812 25-30 mm/yr, 1680

Co

ulo

mb

S

tres

s C

ha

ng

e (

ba

rs)

Distance East of Palmdale (km)

Pre

dic

ted

Sli

p

(cm

)

200180160140120100806040200-20-400

10

20

30

40

50

60

Removes 6.2<M<6.6

Load

Adds Load Equivalent to

5.7<M<6.4

San B

ernard

ino

Indio

Bom

bay

Bea

ch

Slip Needed to RelieveShear Stress Changes

Adds 6.2<M<6.5

Load

Palm

Spri

ngs

Thre

e P

oin

ts

ImmediateStress Changes Adds load

Removes load

Adds load

Removes loadLong-TermStress Changes

200180160140120100806040200-20-40

Coulomb Stress Changeon the San Andreasat 6.25 km depth

-4

0

4

8

µ = 0.75

µ = 0.0

EW

Coulomb Stress Changeon the San Andreas

Fault (for µ = 0.4)

-4

0

4

8

Long-Term(plate)

Immediate(halfspace)

EW

12

Fig. 14 ( ! UP)

Les séismes de Joshua Tree, Landers et Big Bear ont

augmenté !C sur certaines parties de la faille de SA et l'ont diminué sur d'autres

1906

1857Où est une faille à l'intérieur

de son cycle sismique ?

Pas d’accélération visible de la production d’énergie

Pas de séisme proche ?

Accélération de la production d’énergie

Séisme proche ?

1970 1980 1990 20001960

1.0 108

5.0 107

Californie du Nord

aléatoire

non aléatoire

1960 1970 1980 1990 2000

6.0 108

4.0 108

2.0 108

Californie du Sud

aléatoire

non aléatoire

(Bowman et King, 2001)

Comment varie la production d'énergie sismique avant un séisme ?

(Bowman et King, 2001)

Outil de prévision opérationnel ? Pas encore...

La production d’énergie sismique régionale a accéléré avant le séisme

(mais on l'a mesurée après...)

1992 1994 1996 1998 2000 2002 2004

Cum

ula

tiv

e B

enio

ff s

train

0

1.0 106

2.0 106

3.0 106

4.0 106

5.0 106

Year

(Feuillet et King, 2004)